1. A University of Texas Modeling Consortium says that there is an 89 percent chance that we passed the peak in daily deaths. That seems right to me.
Also, that model predicts a sharper decline in New Jersey than in California, which is consistent with the heavy in, heavy out model.
Pointer embedded in a post from Tyler Cowen.
2. The 3DDRR was at 1.15 and excluding New York it was 1.18 The tendency has been for Tuesday to be a peak day for reporting deaths (the Texas people are betting that last Tuesday was the peak), so I am curious to see what tomorrow brings.
An alternative, and to me more reasonable scenario, is presented here. Short version: lockdown can’t continue->exponential spread resumes->society tries and largely fails to find a reasonable compromise between halting the spread of the virus and not destroying the economy. Poor increasingly desperate; riots likely.
At the risk of sounding heartless, from the point of view of public policy, isn’t the number of hospitalizations more important than the number of deaths? Given that no one is predicting millions of deaths anymore, the only conceivable justification for the harm we are doing to ourselves through the lockdown is avoiding a complete breakdown of the health care system. IMHO, anyway.
I really enjoy the ASK blog. But perhaps ASK is barking up the wrong tree with this 3DDRR stuff.
A new study on LA County, from USC–
“221,000 to 442,000 adults” in LA County have COVID-19 antibodies in their blood. No one knows how many additional children have antibodies to COVID-19.
There are 617 COVID-19 related deaths in the county, usually people with co-morbidities.
That’s a 0.01% to 0.03% death rate from COVID-19—but even that measures only adults. The death rate would sink if children were added into the mix. So, deaths are in line with a seasonal flu that was “heavy in,” to quote ASK.
No doubt the USC study will now be attacked or defended on its methodology, because COVID-19 is no longer about policy options, but defending one’s previous positions or even ideologies.
My own take is that hysteria trumped science, and that the macroeconomics profession was totally cowed.
In May 2018, more than 1000 US economists signed a public letter declaring that Trump’s tariffs on China would likely trigger a global Great Depression, just like Smoot-Hawley in the 1930s. The profession easily dismissed any national security or geopolitical concerns behind the tariffs, and absolutely condemned any alternative views about “free trade” and national prosperity.
But then the US not just risked but actively implemented the deep 2020 Depression….and the economists went mute. Even “libertarian” economists, such as Tyler Cowen, assented to lockdowns, though now he is wondering aloud about epidemiologist models.
My conclusion is that free-trade theories are both divine and conventional, and nothing is more powerful than spiritual conventional thinking.
Lockdowns were sacralized as “saving lives” and quickly became convention.
The US macroeconomic craft failed to generate models of lockdown costs involved, including projections of people not paying rents or mortgages for a few seasons….
The antibody tests seem to have a high false positive rate (maybe 1%) because they get confused by antibodies for more common coronaviruses like the common cold. If COVID-19 is still rare, which seems reasonable, nearly all of the positive antibody tests may be false.
https://www.npr.org/sections/health-shots/2020/04/15/834497497/antibody-tests-for-coronavirus-can-miss-the-mark
how then Germans found 15% and in Boston they found 30% on a street quasi-random sample?
Maybe they picked from populations that have recently had colds?
The infection fatality rate that Benjamin Cole suggests isn’t at all consistent with the NYC statistics. If r is the infection fatality rate, n is the number of deaths, and N is the total number of infected people, then r=n/N => N=n/r. Applying these numbers to NYC, where at this writing n=10,344, and using r=0.03%=0.0003, the upper limit suggested by the LA County study, we get N=10,344/0.0003=34,480,000. But the entire population of NYC is only 8,400,000. Even if we assume that every single person in NYC has been infected, and that everyone there who’s going to die of COVID-19 has already done so, we get an infection fatality rate of r=10,344/8,400,000=0.12%, four times as high as the LA County study’s most pessimistic estimate.
who says that IFR will be exactly the same in New York and in California?
One of the things I think is missing from the quantitative discussion is a sense of proportion.
The CDC has early covid mortality estimates 2/1-4/18 here:
https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm
Note that most people who died during that time would have been infected before lockdowns.
If you annualize this, you get 81K deaths for the year.
If you calculate the annualized death rate per 100K, you find that that fatalities per 100K is significantly higher than automobile accidents for _only_ people 75+.
Now, these numbers would obviously grow without lockdowns. But it suggests that a more targeted policy than blanket lockdowns could keep fatalities below our demonstrated risk tolerance.